BGOUG15: JSON support in MySQL 5.7

Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
JSON Support
in MySQL 5.7
Georgi “Joro” Kodinov
Team lead, MySQL server general team

Copyright © 2015, Oracle and/or its affiliates. All rights reserved.
Agenda
2
 The new JSON data type
 Inlined JSON path expressions
 The new JSON functions
 Indexing JSON data

CREATE TABLE employees (data JSON);
INSERT INTO employees VALUES
('{"id": 1, "name": "Jane"}'),
('{"id": 2, "name": "Joe"}');
SELECT * FROM employees;
+-------------------------------------+
| data |
+-------------------------------------+
| {"id": 1, "name": "Jane"} |
| {"id": 2, "name": "Joe"} |
+-------------------------------------+
• Validation on INSERT
• No reparsing on SELECT
• Dictionary of fields
• Fields are sorted
• Can compare JSON/SQL
• Can convert JSON/SQL
• Supports all native JSON
datatypes
• Also supports date, time,
timestamp etc.
3
The New JSON Datatype

JSON vs TEXT columns
Pros Cons
JSON
• Validate once
• Fast access
• Can update in-place
• Slower to insert
• Unreadable as is
• Sets certain limitations on JSON
TEXT
• Fast to insert
• Human readable
• Requires manual validation
• Requires manual parsing
• Harder to update
4

Beware: SQL vs JSON comparisons !
5
SQL JSON
create table t1 (data json); create table t2 (
id integer,
data varchar(20));
insert into t1 values
('{ "id": 1, "data": "1" }'),
('{ "id": 2, "data": "3" }');
insert into t2 values
(1, '1'),
(2, '3');
select count(*) from t1 where
data->'$.id' = data->'$.data';
select count(*) from t2 where
id = data;
0 rows ! 1 row !

Agenda
6

Inlined JSON Path Expressions
• <field>->'<JSON path expression>'
e.g. data->'$.some.key[3].from.doc'
• Syntax sugar over JSON_EXTRACT function
• SELECT * FROM employees WHERE data->'$.id'= 2;
• ALTER … ADD COLUMN id INT AS (data->'$.id') …
• CREATE VIEW .. AS SELECT data->'$.id', data->'$.name' FROM …
• UPDATE employees SET data->'$.name'=‘John' WHERE …
Not
yet!

Limitations of Inlined JSON Path Expressions
Inlined JSON path JSON_EXTRACT()
Data source Field Any JSON value
Path expression SQL Constant SQL Expression
# of expressions One Multiple

Supported JSON Paths
[[[database.]table.]column]$<path spec>
Expression Example
[ [ [database.] table.] field]$ db.phonebook.data$
$ Current document’s root
$.identifier $.user.address.street
[array] $.user.addresses[2].street
.* and [*] $.user.addresses[*].street
** $.user**.phone
Not yet!

Agenda
10

New functions to handle JSON data
Example Result
SELECT JSON_VALID('{ "a":1 }'); 1
SELECT JSON_TYPE('[ 1, 2, 3 ]'); ARRAY
SELECT JSON_KEYS('{ "a":1, "b": 2 }'); ["a", "b"]
SELECT JSON_LENGTH('[ 1, 2, 3 ]'); 3
SELECT JSON_DEPTH('{ "a":{ "c": 1 }, "b": 2 }'); 3
11
Information

Example Result
SELECT JSON_REMOVE('{ "a":1, "b": 2 }', '$.a'); {"b": 2}
SELECT JSON_ARRAY_APPEND('[1,[2,3],4]', '$[1]', 5); [1, [2, 3, 5], 4]
SELECT JSON_SET('{ "a":1 }', '$.c', 3); {"a": 1, “c": 3}
SELECT JSON_INSERT('{ "a":1 }', '$.b', 4); {"a": 1, "b": 4}
SELECT JSON_REPLACE('{ "a":1, "b": 2 }', '$.b', 3); {"a": 1, "b": 3}
SELECT JSON_MERGE('{ "a": 1 }', '{"b":2}'); {"a": 1, "b": 2}
SELECT JSON_UNQUOTE('"abc"'); abc
12
Modification

Example Result
SELECT JSON_ARRAY(1, '2', null, true); [1, "2", null, true]
SELECT JSON_OBJECT(1, 2, '3', true); {"1": 2, "3": true}
SELECT JSON_QUOTE('"null"'); ""null""
13
Create JSON objects

Example Result
SELECT JSON_CONTAINS_PATH(
'{ "a":{ "c": 1 }, "b": 2 }', 'one', '$.a.c');
1
SELECT JSON_CONTAINS( '{"a": 1, "b": "2" }', '1', '$.a'); 1
SELECT JSON_EXTRACT('{"a": 1, "n": { "b": 2}}', '$.n'); {"b": 2}
SELECT JSON_SEARCH( '{"a": "1", "b": "2" }', 'one', 2); "$.b"
14
Search in JSON data

15
Further reading: http://dev.mysql.com/doc/refman/5.7/en/

Agenda
16

Indexing JSON Data
• Use Functional Indexes
– Both STORED and VIRTUAL types are supported
• Examples:
– CREATE TABLE t1 (
data JSON,
id INTEGER AS (JSON_EXTRACT(data,"$.id")) STORED,
PRIMARY KEY(id));
– CREATE TABLE t2 (
data JSON,
id INTEGER AS (JSON_EXTRACT(data,"$.id")) VIRTUAL,
KEY(id));
17

Indexing JSON: STORED vs VIRTUAL columns
Pros Cons
STORED
• Can be primary key too
• All index types supported
• Looks like a normal field
• Slow ALTER TABLE
• Takes space on disk
VIRTUAL
• Instant ALTER TABLE
• Faster INSERT
• Looks like a normal field
• Secondary key only
• BTREE index only
18

Or this way ….
20

Or maybe this way ?
21
ALTER TABLE features ADD feature_type VARCHAR(30) AS (feature->"$.type") VIRTUAL;
Query OK, 0 rows affected (0.01 sec)
Records: 0 Duplicates: 0 Warnings: 0
ALTER TABLE features ADD INDEX (feature_type);
Query OK, 0 rows affected (0.73 sec)
Records: 0 Duplicates: 0 Warnings: 0
SELECT DISTINCT feature_type FROM features;
+-------------------+
| feature_type |
+-------------------+
| "Feature" |
+-------------------+
1 row in set (0.06 sec)
From table scan on 206K documents to index scan on 206K materialized values
Down from 1.25s !
Meta data change only (FAST).
Does not need to touch table.
Online CREATE INDEX !
No rows were
modified.

Roadmap
• Online alter for virtual columns
• Advanced JSON functions
• In-place update of JSON/BLOB
• Full text and GIS index on virtual columns
• Improved performance through condition pushdown
22

Questions ?
@gkodinov, georgi.kodinov@oracle.com if you forget !
23

Safe Harbor Statement
The preceding is intended to outline our general product direction. It is intended for
information purposes only, and may not be incorporated into any contract. It is not a
commitment to deliver any material, code, or functionality, and should not be relied upon
in making purchasing decisions. The development, release, and timing of any features or
functionality described for Oracle’s products remains at the sole discretion of Oracle.
24

BGOUG15: JSON support in MySQL 5.7

More Related Content

What's hot (20)

Viewers also liked (9)

Similar to BGOUG15: JSON support in MySQL 5.7 (20)

Recently uploaded (20)

BGOUG15: JSON support in MySQL 5.7

Editor's Notes